Research On Multi-Objective Optimization Task Offloading Strategy Based On Edge Computing

With the rapid development of Internet of Things technology and the popularization of 5G network communication technology,the number of smart terminal devices has increased sharply,and network traffic data has also increased exponentially.The traditional cloud computing model has been somewhat weak in the face of this situation.It is against this background that edge computing was born.By sinking the resources of the cloud computing center to the edge devices of the network,it provides computing and storage services for smart terminal devices nearby,thereby shortening the processing delay and energy consumption of tasks and ensuring the quality of user services.In view of the limitations of multi-objective optimization problems such as delay and energy consumption in the current edge computing task offloading process,this thesis studies the multi-objective optimization task offloading strategy of edge computing.The specific work content is as follows:(1)Aiming at the limitations of the traditional heuristic algorithm in solving the problem of task unloading strategy,this thesis proposes a hybrid particle swarm optimization algorithm,which absorbs the respective advantages of genetic algorithm and particle swarm optimization algorithm.A better task offloading strategy can be achieved on multi-objective optimization problems such as time delay and energy consumption.First,this thesis constructs a multi-terminal multi-edge server task offloading scenario,and then establishes a multi-objective optimization model for factors such as delay,energy consumption,cost,and load balancing in the task offloading process.In the process of solving the model,the inertia weight update method of the particle swarm optimization algorithm is improved,and the ideas of genetic algorithm selection,crossover and mutation are incorporated to avoid the problem that the particles are easy to fall into the local optimal solution,and at the same time improve the convergence of the algorithm speed.Finally,the superiority of the algorithm in solving the task offloading strategy problem is verified by experiments.The experimental results show that the hybrid particle swarm optimization algorithm proposed in this thesis is superior to the traditional heuristic algorithm in terms of time delay and energy consumption,and the average time delay and energy consumption are reduced by 12.6% and 10.3% respectively.(2)Aiming at the divisibility of tasks in the offloading process,this thesis proposes a fine-grained task offloading strategy.This strategy aims to divide each task and subdivide the task into multiple subtasks with interdependence.First of all,this thesis builds a task subdivision model.By analyzing the serial and parallel sequences between each subtask,the subtasks of the same layer are parallelly offloaded to the edge server for execution,and at the same time,the subtasks with dependencies are maximized and merged to the same edge server.on processing.Then a multi-objective optimization model is established for factors such as delay,energy consumption,cost and load balance in the process of task offloading.Finally,the MOEA/D algorithm is used to solve the model to obtain a set of candidate task offloading strategies,and the optimal task offloading strategy is screened out using the offloading decision matrix.The simulation results show that the task fine-grained offloading strategy proposed in this thesis is better than the task coarse-grained offloading strategy in all indicators,and the average delay and energy consumption are reduced by 11.1% and 9.7% respectively.